A New Blind Watermark Embedding Model: Spiral Updated Rider Optimization Algorithm. (1st February 2021)

• Record Type: Journal Article
• Authors: Narasimhulu, C Venkata
• Is Part Of: Computer journal. Volume 65:Number 6(2022)
• Page Start: 1365
• Page End: 1385
• Publication Date: 2021-02-01
• Subjects: multimedia watermarking -- Haar wavelet -- watermark embedding -- watermark detection -- optimization; Computers -- Periodicals; 005.1
• DOI: 10.1093/comjnl/bxaa185
• Languages: English
• ISSNs: 0010-4620

Abstract

Abstract: In this era, the exchanging of multimedia content has become much comfortable. Since the distribution and the illegal copying of digital multimedia content have become much easier, an author's intellectual property copyrights have suffered from the violations that led to many damage to their benefits in many applications. This paper intends to propose a novel digital watermarking model for multimedia copyright protection that is based on 3D wavelet transformation. Initially, histogram-based motion frame extraction is conducted on the color video for which keyframes are extracted based on the extracted motion frames. To embed the watermark in the keyframes, 3D wavelet coefficients are selected followed by applying the spread spectrum technique. To conduct precise and effective embedding, it is planned to introduce fine-tuning aspects in the embedding process. As the intensity factor plays a major role in the embedding process, this paper tunes it rather than selecting randomly. Moreover, the Mersenne-Twister algorithm is known for spreading the power spectrum of watermark data to keep invisible. Hence, it is also intended to tune the seed factor of the respective algorithm to further effectuate the spreading process. These tuning processes can be solved as an optimization problem for which this paper proposes a new hybrid optimization algorithm named Spiral Updated Rider Optimization Algorithm (SU-ROA). SU-ROA is a hybrid variant of the Rider Optimization Algorithm (ROA) and Whale Optimization Algorithm (WOA). The performance of the proposed work is compared and proved over the state-of-the-art models in terms of both embedding and extracting effectiveness.